The present invention relates to a data conversion method, and more particularly to a data conversion method that is suitable for the use in a large scale database, and that enables a reduction in load of a data conversion server as well as an easy design of a conversion program.
In recent years, database systems are more and more being developed, and are therefore becoming systems that are indispensable to company activities. A database system is not only used in one computer system, but it is often exported to other computer systems, or partial data of its table is converted before the table is used. In particular, as a scale of a database system becomes larger, such data conversion is more often required.
In such a case, for the purpose of applying certain table data to various kinds of uses, it is necessary to modify an original table so as to create a plurality of application-specific tables, and then to use each of the tables in each system. Therefore, a data conversion program for creating those tables is required. The data conversion program obtains an input and table data from the storage device in which the data is stored, and executes data conversion required for the use of a system, and then writes data after the conversion to a storage device as table data that is an output. For example, in Japanese Patent Application Laid-Open No. 2001-22764, the technology for converting a data file in a tabular format into product information database is disclosed.
Here, field values of a table in which input data is stored can be classified into two kinds of data: field data which needs to be rewritten, and field data which does not need to be rewritten. The field data which needs to be rewritten is field data which needs to be rewritten for a different use because this field data differs depending on the use of the system. On the other hand, the field data which does not need to be rewritten is field data which does not need to be rewritten because this field data of the original table is used in common. These classifications differ depending on the use of the system to which a table is output. Additionally, how to classify the field values is also flexible depending on the use.
For these reasons, in a data conversion job of the conventional data conversion program, irrespective of the classifications of fields in a table, target table data to be converted is entirely read into a data conversion server in which a data conversion program operates. Then, only field data for which rewriting is specified is converted. After having executed the data conversion in the specified fields of the obtained table, the converted data is written as output table data. According to another data conversion method using a data conversion job, a table as an input is first copied, and then only required field data of the copied table is converted.
Here, the data conversion job is described as a job definition in which a model, an obtaining method, and a conversion method, of input/output data, and the origin, and processing, of data are described. The job definition is interpreted and executed by a job execution engine.
In the data conversion method according to the prior art, all table data to be converted were read into the data conversion server used for operating the data conversion program. Under such circumstances, because I/O load and all of the data read into the data conversion server required CPU processing, data which did not need to be converted also caused a CPU load. This overhead produced a problem of treatment performance. Moreover, in the method for designing a data conversion job in which only required data is converted, a designer of a data conversion job was required to have extra skills such as knowledge about the storage device for storing the table data, how to judge field data, and how to manage the processing order of the field data.
Thus, the conventional method had the problem of treatment performance because of the load being placed on the data conversion server, or the problem of high skills which a data conversion job designer is required to have.
The present invention is devised to solve the above-mentioned problems. An object of the present invention is to provide a method that can reduce a load of a data conversion server when converting data of a database, and that enables a designer of a data conversion job to easily design the data conversion job.